This action might not be possible to undo. Are you sure you want to continue?
•Tuesday, November 13, 2001
•Laura Squier •Technical Consultant •email@example.com
• What Data Mining IS and IS NOT • Steps in the Data Mining Process – CRISP-DM – Explanation of Models – Examples of Data Mining Applications • Questions
The Evolution of Data Analysis
Evolutionary Step Business Question Enabling Technologies Product Providers Characteristics Data Collection (1960s) "What was my total Computers, tapes, revenue in the last disks five years?" "What were unit sales in New England last March?" Relational databases (RDBMS), Structured Query Language (SQL), ODBC On-line analytic processing (OLAP), multidimensional databases, data warehouses Advanced algorithms, multiprocessor computers, massive databases IBM, CDC Retrospective, static data delivery
Data Access (1980s)
Oracle, Sybase, Informix, IBM, Microsoft
Retrospective, dynamic data delivery at record level
Data Warehousing & Decision Support (1990s)
"What were unit sales in New England last March? Drill down to Boston." "What’s likely to happen to Boston unit sales next month? Why?"
SPSS, Comshare, Retrospective, Arbor, Cognos, dynamic data Microstrategy,NCR delivery at multiple levels
Data Mining (Emerging Today)
SPSS/Clementine, Lockheed, IBM, SGI, SAS, NCR, Oracle, numerous startups
Prospective, proactive information delivery
Results of Data Mining Include: • Forecasting what may happen in the future • Classifying people or things into groups by recognizing patterns • Clustering people or things into groups based on their attributes • Associating what events are likely to occur together • Sequencing what events are likely to lead to later events .
Data mining is not •Brute-force crunching of bulk data •“Blind” application of algorithms •Going to find relationships where none exist •Presenting data in different ways •A database intensive task •A difficult to understand technology requiring an advanced degree in computer science .
interactive process which leverages analysis technologies and computing power •A group of techniques that find relationships that have not previously been discovered •Not reliant on an existing database •A relatively easy task that requires knowledge of the business problem/subject matter expertise .Data Mining Is •A hot buzzword for a class of techniques that find patterns in data •A user-centric.
Data Mining versus OLAP •OLAP .On-line Analytical Processing – Provides you with a very good view of what is happening. but can not predict what will happen in the future or why it is happening .
•Data Analysis •Data Mining – Originally developed to act – Tests for statistical correctness of models as expert systems to solve • Are statistical problems assumptions of models – Less interested in the correct? mechanics of the – Eg Is the R-Square technique good? – If it makes sense then let‟s – Hypothesis testing use it • Is the relationship – Does not require significant? assumptions to be made about data – Use a t-test to validate significance – Can find patterns in very – Tends to rely on sampling large amounts of data – Techniques are not – Requires understanding optimised for large amounts of data and business of data problem – Requires strong statistical skills Data Mining Versus Statistical Analysis .
waste and selling.Examples of What People are Doing with Data Mining: •Fraud/Non-Compliance Anomaly detection •Recruiting/Attracting customers •Maximizing – Isolate the factors that profitability (cross lead to fraud. identifying abuse profitable customers) – Target auditing and •Service Delivery and investigative efforts more Customer Retention effectively – Build profiles of customers likely to use which services •Credit/Risk Scoring •Intrusion detection •Parts failure prediction •Web Mining .
How Can We Do Data Mining? By Utilizing the CRISPDM Methodology – a standard process – existing data – software technologies – situational expertise .
•Framework for recording experience – Allows projects to be replicated •Aid to project planning and management •“Comfort factor” for new adopters – Demonstrates maturity of Data Mining – Reduces dependency on “stars” .Why Should There be a Standard Process? The data mining process must be reliable and repeatable by people with little data mining background.
Daimler-Benz. Magnify. … – End Users .Cap Gemini. AirTouch.Process Standardization CRISP-DM: • • • • • CRoss Industry Standard Process for Data Mining Initiative launched Sept. SAS. NCR. .BT. Syllogic.SPSS. NCR.1996 SPSS/ISL. Lloyds Bank. – System Suppliers / consultants . Experian. ABB.. OHRA Funding from European commission Over 200 members of the CRISP-DM SIG worldwide – DM Vendors . .. . SGI. Deloitte & Touche. ICL Retail.. Data Distilleries. IBM.
CRISP-DM •Non-proprietary •Application/Industry neutral •Tool neutral •Focus on business issues – As well as technical analysis •Framework for guidance •Experience base – Templates for Analysis .
The CRISPDM Process Model .
Why CRISP-DM? •The data mining process must be reliable and repeatable by people with little data mining skills •CRISP-DM provides a uniform framework for –guidelines –experience documentation •CRISP-DM is flexible to account for differences –Different business/agency problems –Different data .
Phases and Tasks Business Understanding Data Understanding Data Preparation Modeling Evaluation Deployment Determine Business Objectives Collect Initial Data Background Business Objectives Business Success Criteria Initial Data Collection Report Describe Data Data Set Data Set Description Select Data Select Modeling Technique Evaluate Results Data Description Report Explore Data Rationale for Inclusion / Exclusion Clean Data Modeling Technique Modeling Assumptions Generate Test Design Situation Assessment Test Design Assessment of Data Mining Results w. Assumptions.r. and Constraints Risks and Contingencies Terminology Costs and Benefits Determine Data Mining Goal Data Exploration Report Verify Data Quality Data Cleaning Report Construct Data Build Model Data Quality Report Derived Attributes Generated Records Integrate Data Parameter Settings Models Model Description Assess Model Review of Process Determine Next Steps Produce Final Report List of Possible Actions Decision Merged Data Format Data Model Assessment Revised Parameter Settings Experience Documentation Data Mining Goals Data Mining Success Criteria Reformatted Data Produce Project Plan Project Plan Initial Asessment of Tools and Techniques .t. Business Success Criteria Approved Models Review Process Plan Deployment Deployment Plan Plan Monitoring and Maintenance Monitoring and Maintenance Plan Final Report Final Presentation Review Project Inventory of Resources Requirements.
Phases in the DM Process: CRISP-DM .
Phases in the DM Process (1 & 2) •Business Understanding: – Statement of Business Objective – Statement of Data •Data Understanding Mining objective – Explore the data and – Statement of Success verify the quality Criteria – Find outliers .
aggregation level.create new variables .Phases in the DM Process (3) • Data preparation: – Takes usually over 90% of our time • Collection • Assessment • Consolidation and Cleaning • Data selection – outliers? – table links. etc – active role in ignoring noncontributory data? – Use of samples – visualization tools • Transformations . missing values.
different for supervised and unsupervised learning • May model for either description or prediction .Phases in the DM Process (4) • Model building – Selection of the modeling techniques is based upon the data mining objective – Modeling is an iterative process .
logistic regression) . GLM) GRI – Classification algorithm predict symbolic outcome): CHAID.0 (discriminant analysis. algorithms: apriori.Types of Models •Prediction Models for •Descriptive Models for Predicting and Grouping and Finding Classifying Associations – Regression algorithms – Clustering/Grouping (predict numeric algorithms: Koutcome): neural means. C5. CART – Association (OLS regression. rule induction. Kohonen networks.
Neural Network Input layer Hidden layer Output .
Neural Networks • Description – Difficult interpretation – Tends to „overfit‟ the data – Extensive amount of training time – A lot of data preparation – Works with all data types .
% n Bad 52.92) 158 Old ( > 35) Cat.Rule Induction •Description – Produces decision trees: • income < $40K – job > 5 yrs then good risk – job < 5 yrs then bad risk – high debt then bad risk – low debt then good risk Weekly pay Cat.08) 165 Age Categorical P-value=0. % n Bad 15.0388. % n Bad 0.75) 109 • income > $40K – Or Rule Sets: • Rule #1 for good risk: – if income > $40K – if low debt • Rule #2 for good risk: – if income < $40K – if job > 5 years . % Bad 0. df=1 Young (< 25). Chi-square=179.98 24 Good 51. % n Bad 90.92) 158 Age Categorical P-value=0.00 Total (2. % n Bad 58.67 143 Good 13.01 168 Good 47. % n Bad 86.Middle (25-35) Cat.7255.17) n 0 7 7 Credit ranking (1=default) Cat.0000. df=1 Young (< 25) Cat.0016. df=1 Management.0000.00) 323 Paid Weekly/Monthly P-value=0.02 25 Total (15.49 15 Total (48.Clerical Cat.6665. % Bad 0.54 24 Good 41.00 Good 100. % n Bad 48.51 143 Good 9.00 Good 100. Chi-square=12.Old ( > 35) Cat.82 25 Good 84.1113.92 1 Good 99.18 133 Total (48.46 17 Total (12.0000. Chi-square=30.00 Total (2. df=1 Monthly salary Cat.69) 41 Middle (25-35).08 108 Total (33.48) n 0 8 8 Professional Cat. Chi-square=58.17) 49 Social Class P-value=0.99 155 Total (100.33 22 Total (51.
Rule Induction Description • Intuitive output • Handles all forms of numeric data. as well as non-numeric (symbolic) data C5 Algorithm a special case of rule induction • Target variable must be symbolic .
Apriori Description • Seeks association rules in dataset • „Market basket‟ analysis • Sequence discovery .
Kohonen Network Description • unsupervised • seeks to describe dataset in terms of natural clusters of cases .
g. easy or hard depends on algorithm . mean error rate with regression models – Interpretation of model: important or not.Phases in the DM Process (5) • Model Evaluation – Evaluation of model: how well it performed on test data – Methods and criteria depend on model type: • e.. coincidence matrix with classification models.
Phases in the DM Process (6) •Deployment – Determine how the results need to be utilized – Who needs to use them? – How often do they need to be used •Deploy Data Mining results by: – Scoring a database – Utilizing results as business rules – interactive scoring on-line .
Specific Data Mining Applications: .
The US Internal Revenue Service needed to improve customer service and. Scheduled its workforce to provide faster. .... more accurate answers to questions..What data mining has done for.
..What data mining has done for. The US Drug Enforcement Agency needed to be more effective in their drug “busts” and analyzed suspects‟ cell phone usage to focus investigations..
.. Reduced direct mail costs by 30% while garnering 95% of the campaign‟s revenue..What data mining has done for.. HSBC need to cross-sell more effectively by identifying profiles that would be interested in higher yielding investments and..
• By using the CRISP-DM methodology. and valid results.Final Comments • Data Mining can be utilized in any organization that needs to find patterns or relationships in their data. repeatable. . analysts can have a reasonable level of assurance that their Data Mining efforts will render useful.
This action might not be possible to undo. Are you sure you want to continue?
We've moved you to where you read on your other device.
Get the full title to continue listening from where you left off, or restart the preview.